Identify the bugs that are causing tests to fail when transform is run on the auditboard app

[mansona]

Background

As we are still in the experimental phase of the project, there are still a few rough edges to some of the tools that we are producing. The intl-template-transform works fantastically well in most cases and we have been able to successfully run it against the Auditboard app, with the app booting and most tests are passing 🎉

You can see the PR here: https://github.com/soxhub/soxhub-client/pull/11398

Remaining work

I mentioned that most tests are passing, which means that some are failing 😞 you can see by looking at the checks of the PR above that there are a number of different test batches that are failing. Some are reasonably easy to identify like https://github.com/soxhub/soxhub-client/runs/4119135755?check_suite_focus=true#step:6:31 which tells you the exact error and we know what we need to fix, other test failures are a little bit more complicated because it seems like they are crashing the test runner 🤔 https://github.com/soxhub/soxhub-client/runs/4119135648?check_suite_focus=true

I have attempted to run the whole test suite locally to identify the issues but that seems to crash the browser too, so we will need to follow a different approach to identifying the failing tests.

Recommended approach

As each of the tests are run in batches e.g. percy exec ember exam --filter='!API | ' --path='test-dist' --split=10 --partition=$PARTITION we should be able to recreate the batches that are failing locally, and we should be able to list the test names that are passing or failing.

This may need to be a manual process of elimination because of the fact that it crashes your browser when run locally.

Deliverable

The most important deliverable of this task is to have a list of exactly which tests are failing and why. We don't need to fix the transform to avoid those issues (unless it's trivial to do so) but it would be worth having sample code that could be added to #7 for future work.

[BobrImperator]

This issue seems to be referring to tests failing if the transform is ran for the whole app in one big bulk.
This isn't how we want to migrate translations - we want to do it in smaller chunks so any issues that transform could cause are easier to detect.

The issue can be closed IMO.
(Hopefully) will revisit and close.